DOI: https://doi.org/10.21203/rs.2.19229/v1
Background: Internet Addiction is a newly emerging issue for adolescents globally. Previous literature suggested a tangled relationship between Internet Addiction, depressive mood and lower subjective and objective ratings in Quality of Life. However, the independent relationship between Internet Addiction and Quality of Life was less discussed.
Method: A cross-sectional study was conducted in Shenzhen adolescents to address the independent associations between Internet Addiction and 4 domains of Quality of Life (physical, psychological, social relation and environmental domains), in order to further guide the prevention and intervention for IA. Associations were calculated using linear regression models before and after the adjustment for the existence of depressive mood and confounding demographic factors.
Result: The prevalence of Internet Addiction in Shenzhen was 23.2%. Engaging with Internet Addiction resulted in significant decrease in Quality of Life score in physical domain (B = -0.87, p<.001) and psychological domain (B = -0.40, p = 0.011) as well as insignificant decrease in social relation domain (B = -0.36, p = 0.063) and increase in environmental domain (B = 0.02, p = 0.906).
Conclusion: Although Quality of Life is considered to be related to many factors, Internet Addiction on its own had a significant impact on lower subjective life satisfaction overall, especially in psychical and psychological wellbeing. Key Words:Internet Addiction, Quality of Life, Adolescents, Depressive mood, Chinese
Within last two decades, our life has changed rapidly as the so-called era of Internet arrives. Due to the escalation of availability to Internet, the proportion of people who use Internet in the world raised dramatically from less than 1% in 1995 to over 46% in 2016 [1] and over 70% of youth were online in 2017 [2]. In addition to loads of benefits that Internet brings, expanding in the accessibility to Internet may also increase the susceptibility to Internet Addition (IA) [3]. IA is defined as incapability to inhibit internet use, which eventually results in deficits of one’s daily life and psychological well-being [4]. It is characterized by symptoms including: (1) need for extending online time (tolerance), (2) withdrawal symptoms, (3) inadequate control on Internet use, (4) continuation of Internet use disregarding to problem awareness, (5) excessive online time, (6) negative consequences [5, 6]. A bunch of terms were used in literature to describe this condition, such as Pathological Internet Use [7], Internet Abuse [8], Internet Use Disorder, Problematic Internet Use [9] and etc. In the third section of the Fifth Edition of Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [10], which proposed conditions for future study, Internet Gaming Disorder is included as a potential diagnosis for further discussion. Given that Internet Addiction is the most commonly used term [11, 12], it was then adopted in this current paper for consistency.
Since 2000, large scale studies on the prevalence of IA in youth population found an overt variation across studies, ranging from 0.8–26.7% [13]. In 2014, Cheng and Li [14] pointed out in their systematic review that global occurrence of IA was 6.0% across 31 nations. Although huge discrepancies in settings, population, year of investigation and assessment tools raised concerns, there was a remarkable high rate of IA in youth. The impact of IA on adolescents’ life and functioning, therefore, is valuable to discuss.
IA is associated not only with people’s subjective views on their quality of life (QoL) (e.g. physical health and psychological wellbeing), but also objective assessment on QoL (e.g. quality of environmental conditions) [14]. Previous literature has illustrated that IA might cause deterioration of social and emotional competence of a person, which eventually leads to a poorer satisfaction of life [15, 16]. In 1998, Kraut and colleagues [17] proposed a ‘Displacement Theory’, stating that people’s social development would be undermined by decreased time spent in real life socialization, as excessive time spent online. Social relationships are crucial to keep positive emotions and well-being. Heavy Internet users, who donate excessive time to online relationship or extramarital affairs, are more likely to experience family problems and social difficulties in real world [4, 18]. IA is thus supposed to affect people’s real-life social quality and well-being. However, contradictions in research can be seen. Some research suggested positive effect of internet use on foster closer interpersonal relationship [19, 20].
Moreover, IA has also been reported to be significantly related to physical problems and mental health issues [21, 22]. People who are suspicious to have an Internet gaming disorder often reported a lower level of physical activity [23]. But for the consensus in most research, some large scale studies reported that a marginally higher level of physical health was revealed in people with IA [23] .
Psychological well-being of a person is another domain that is highly correlated with IA. A high co-morbidity rate of IA and other mental disorders was reported in previous research, including Attention Deficit Hyperactivity Disorder (ADHD), anxiety and depressive disorder [23, 24]. Lots of studies aimed to illustrate the relationship between IA and depressive mood in different populations. In Chinese Han and Tibetan adolescents, those met the criteria for IA claimed to be more likely to meet the criteria for depressive mood and to have a lower rating in subjective psychological QoL [25]. Also, studies attempted to understand the causal relationship between depressive mood and IA. Yu and Shek [6] did a follow-up study and suggested a causal relationship between IA at baseline and later onset depressive mood, resulting in a poorer life satisfaction in terms of mental wellbeing. On the other hand, IA is proposed to be dependent on individual’s mood. In the Mood Enhancement Hypothesis, people experiencing negative emotion may engage to internet use, in order to alleviate stress and boost their mood [26]. This is consistent with Person-Affect-Cognition-Executive (I-PACE) model. I-PACE posit that the development and maintenance of IA is predisposed by biopsychological constitutions including genes and personalities, which is mediated by affective and cognitive responses and confounded by coping style and executive functions, resulting in negative consequence in daily life and deteriorating psychological health of people eventually [27]. Thus, a tangled relationship among IA, depressive mood and self-perceived QoL was suggested and deserved further research.
Furthermore, more attention has been drawn to health outcome. As a significant component in measuring people’s health outcome, QoL and its relation to IA is a hotly investigated field in the last decade. Depressive mood is highly prevalent in adolescent, which is also a significant factor concerning one’s life satisfaction. Depressive mood or depression, as aforementioned, may trigger IA at first and eventually exacerbate as a consequence of IA in I-PACE model. Depression is recognized as the one of the factors that leads to increase in Disability-adjusted life year (DAILY) and decrease in Healthy life expectancy (HLE) worldwide [28]. Previous studies explored the relationship between IA and QoL, indicating risk factors (including depressive mood) of IA [25, 29]. However, the pure relationship between IA and different domains of QoL, independent of other significant factors, including depressive mood, is less discussed. In order to promoting the invention on IA, it is of importance to illustrate whether IA on its own has a significant relation to QoL independent from depressive mood. This would eventually help in guiding population-based intervention in maintaining a healthy life satisfaction.
Additionally, understanding the pattern of IA in different population is argued to be significant to develop preventive and treatment plans [6]. Also, different sociocultural factors can have impact on the pattern of IA in an area [25, 30]. Shenzhen is recognized as the earliest Special Economic Zone in China set up in 1980s. Given its openness to worldwide and its pioneer position in IT industry in China, people in Shenzhen have great exposure to Internet. Also, Shenzhen is one of the major cities in the China Great Bay Area (GBA). However, fewer epidemiology studies investigating IA prevalence and association to psychosocial factors were conducted in Shenzhen in comparison to Hong Kong, another major city in GBA. Thus, it is necessary to explore the pattern of IA and its related factors and influences in Shenzhen.
The current study aimed to firstly reveal the prevalence of IA and depressive mood in Shenzhen adolescents. It will then identify the relationship between IA and QoL, both overall and separate domains of QoL, and discuss whether there is a relationship between them, independent from depressive mood and other significant demographic factors.
The sample was collected in Shenzhen Middle schools in June and October 2017. Eleven middle schools from two districts of Shenzhen were involved. In total, eight classes of each School Year (Grade 1 to Grade 8) were randomly selected to join in the current cross-sectional study. All students in the selected classes were invited to complete a self-designed questionnaire anonymously. These questionnaires, together with consent forms, were distributed to the participants by psychological teachers at these schools and were collected after completion. Anyone who refused to sign the consent was excluded. In total, 2200 questionnaires were sent out, 2148 eligible questionnaires were collected. The representativeness of the sample was shown through the commonness of demographic information of these students to the current middle school students in Shenzhen City.
A self-designed questionnaire was sent out to collect basic sociodemographic information of subjects. Information including age, gender, current grade, academic performance and self-perceived academic pressure, number of children in their family, family social economic status, parental information (occupations, educational level, marital status), relationship with family, schoolmates and teachers, self-perceived health and body weights, exercise level and sleep quality were collected.
Internet Addiction Test (IAT), a 20-item self-rated scale, was adopted to examine the presence and severity of IA. IAT was developed by Young [4, 31] and was validated in many languages, including Chinese [32, 33]. Each item was rated on a scale of 1 to 5. A total score of over 49 was recognized as moderate to severe dependence, which was defined as “having IA” [31]. This scale was widely utilized previously in China as well and satisfactory psychometric properties was reported (Cronbach’s Alpha equals 0.713) [6, 25].
Centre for Epidemiologic studies of Depression Symptom Scale (CES-D), a widely used self-report scale to measure depressive symptomatology in general population, was employed to examine the depressive mood of subjects. This scale was validated in many counties with different age groups [34]. It contains 20 questions. Each question was rated from 0 (“rarely”) to 3 (“always”), resulting in a total score of 0 to 60. The higher the score, the greater depressive mood was indicated. Usually, a score of greater than 15 was reported as “above cut-off”. This was the approximate 80th percentile in the original Community Mental Health Assessment (CMHA) study [34]. The current study used “≥28” as a cut-off line, which selected about 5% of CMHA sample [34], in order to gain a rate that generate a sample closer to clinical sample.
A short version of World Health Organization Quality of Life (WHOQOL-BREF) was used to measure their life quality [35, 36]. It consists of 26 items, to assess four domains including physical health, psychological health, social relationships and environment. A Chinese version was validated and utilized in previous studies [25, 37].
The data was analyzed using IBM SPSS, version 2.1. Firstly, descriptive analysis was carried out to explore the prevalence of depressive emotion, internet addiction and other demographic information in the current sample. The demographic features were compared between the complete sample and the whole sample, to understand whether missing data play a significant role in the current sample. Subsequently, the comparison between students with and without IA was analyzed through independent Sample T tests, chi square tests and Mann Whitney U tests in regards to differences in depressive emotion, QoL evaluation and related demographic factors. Thereafter, the correlations between continuous basic demographic and clinical characteristics (internet addiction, depressive emotion and QoL) were then examined by Pearson’s Correlation for normally distributed variables and Spearman Correlation for non-distributed variables. After that, Analysis of Covariance (ANCOVA) was employed to compare the QoL in four different domains between students with and without IA. In order to exam the association between QoL and IA, univariate regression was performed to understand the association between IA and WHOQoL scores in four different domains. Finally, multivariate regression was performed to explore the independent relationship between IA and QoL, where factors including depressive emotion and other related demographic factors were controlled.
In total, 2,148 students were consented and completed the questionnaire, of which 1,900 students provided eligible data on Internet Addiction Test. Within this sample, 1,725 students completed the scale for QoL (WHOQoL). The sample was further reduced to 1,482 whose data were available for all possible confounding variables. Figure 1 illustrated how the sample was restricted to the final complete sample in a flow chat. The demographic characteristics and clinical features of the complete case sample were summarized in Table 1, with comparison to the whole sample. The age of the complete case sample ranged from 9 to 18 years, with an average of 15.29, significantly different from the average age of the whole sample. Other characteristics of the complete case sample and the whole sample were not significantly different.
Demographic Characteristics | Complete Case Sample (N = 1,482) | Whole Samplea | p | ||||||
---|---|---|---|---|---|---|---|---|---|
N | % | Total N | N | % | |||||
Male | 699 | 47.2 | 2,100 | 1073 | 51.1 | ||||
Grade | 2,131 | ||||||||
Grade 7 | 240 | 16.2 | 363 | 17 | |||||
Grade 8 | 200 | 13.5 | 312 | 14.6 | |||||
Grade 9 | 159 | 10.7 | 308 | 14.5 | |||||
Grade 9+ | 1 | 0.1 | 9 | 0.4 | |||||
Grade 10 | 327 | 22.1 | 408 | 19.1 | |||||
Grade 11 | 282 | 19 | 365 | 17.1 | |||||
Grade 12 | 273 | 18.4 | 366 | 17.2 | |||||
Perceived Health Status | 2,124 | ||||||||
Good | 619 | 41.8 | 970 | 45.7 | |||||
Average | 773 | 52.2 | 1025 | 48.3 | |||||
Poor | 90 | 6.1 | 129 | 6.1 | |||||
Perceived Weight | 2,118 | ||||||||
Underweight | 182 | 12.3 | 262 | 12.4 | |||||
Average | 808 | 54.5 | 1176 | 55.5 | |||||
Overweight | 492 | 33.2 | 680 | 31.7 | |||||
Academic performance (scores) | 2,089 | ||||||||
85 ~ 100 | 223 | 15 | 318 | 15.2 | |||||
75 ~ 85 | 433 | 29.2 | 602 | 28.8 | |||||
60 ~ 75 | 664 | 44.8 | 914 | 43.8 | |||||
lower than 60 | 162 | 10.9 | 255 | 12.2 | |||||
Perceived academic pressure | 2,128 | ||||||||
Light | 71 | 4.8 | 137 | 6.4 | |||||
Average | 890 | 60.1 | 1256 | 59 | |||||
Heavy | 521 | 35.2 | 735 | 34.5 | |||||
Family Social Economic Status | 2,116 | ||||||||
Rich | 237 | 16 | 371 | 17.5 | |||||
Enough | 1178 | 79.5 | 1639 | 77.5 | |||||
Poor | 67 | 4.5 | 106 | 5 | |||||
Parental Marriage Status | 2,116 | ||||||||
Normal | 1339 | 90.4 | 1893 | 89.5 | |||||
Re-marriage | 34 | 2.3 | 55 | 2.6 | |||||
Divorced | 64 | 4.3 | 99 | 4.6 | |||||
Live apart | 27 | 1.8 | 38 | 1.8 | |||||
Widowed | 7 | 0.5 | 14 | 0.7 | |||||
Others | 11 | 0.7 | 17 | 0.8 | |||||
Family Relationship | 2,120 | ||||||||
Good | 1066 | 71.9 | 1512 | 71.3 | |||||
Average | 392 | 26.5 | 554 | 25.8 | |||||
Poor | 24 | 1.6 | 54 | 2.5 | |||||
Relationship with Teacher | 2,126 | ||||||||
Good | 564 | 38.1 | 863 | 40.6 | |||||
Average | 883 | 59.6 | 1190 | 56 | |||||
Poor | 35 | 2.4 | 73 | 3.4 | |||||
Relationship with Classmates | 2,127 | ||||||||
Good | 780 | 52.6 | 1174 | 55.2 | |||||
Average | 670 | 45.2 | 893 | 42 | |||||
Poor | 32 | 2.2 | 60 | 2.8 | |||||
Mean | SD | Total N | Mean | SD | |||||
Age | 15.29 | 1.6 | 2,058 | 15.15 | 1.76 | 0.001 | |||
CES-D Score | 15.18 | 10.03 | 1,971 | 15.1 | 10.19 | 0.764 | |||
WHOQOL Score | 1,910 | ||||||||
Physical Domain | 14.53 | 2.51 | 14.48 | 2.67 | 0.415 | ||||
Psychological Domain | 13.85 | 3.13 | 13.87 | 3.27 | 0.782 | ||||
Social Relationship | 14.9 | 3.58 | 14.87 | 3.73 | 0.746 | ||||
Environmental Domain | 14.26 | 3.37 | 14.24 | 3.08 | 0.819 | ||||
Overall | 14.16 | 3.37 | 14.27 | 3.5 | 0.206 | ||||
Internet Addiction Test Score | 41.09 | 13.8 | 1,900 | 41.44 | 14.29 | 0.331 | |||
N | % | Total N | N | % | |||||
Met the cut-off for Internet Addiction | 344 | 23.2 | 1,900 | 460 | 24.2 | ||||
Met the cut-off for Depressive Emotion | 172 | 11.6 | 1,971 | 231 | 11.7 | ||||
NOTE: a. The whole sample includes both complete case and excluded sample, N = 2,148. Percentages were calculated based on those with the data on the variable, indicated in ‘Total N’ column. |
The comparison of characteristics between subjects with and without IA was conducted by Mann Whitney U tests for factors including age, CES-D score and QoL scores, as these data were skewed. Variables indicating frequencies were compared using Chi Square Tests. The results suggested significant differences in the grade, perceived health status, perceived weight, perceived study pressure, SES, and relationships with family, classmates and teacher between students with and without IA (See Table 2). After controlling confounding variables, the comparison of QoL scores between subjects with and without IA by ANCOVA showed significant differences in physical health domain (F (1, 1,480) = 21.30, p < .001) and in environmental domain (F (1,1480) = 3.91,p = 0.048). People with IA were worse in physical satisfaction but better in environmental satisfaction. However, their QoL scores were not significantly different in social relation domain (F(1, 1480) = 0.21, p = 0.643), psychological domain F(1, 1480) = 0.20, p = 0.658) or overall satisfaction (F(1, 1,480) = 1.01, p = 0.314).
No Internet Addiction (N = 1138) | Internet Addiction (N = 344) | Statistics | |||||||
---|---|---|---|---|---|---|---|---|---|
Demographic Features | Mean | SD | Mean | SD | Z | df | p | ||
Age | 15.25 | 1.61 | 15.41 | 1.56 | -1.57 | 0.118 | |||
N | % | N | % | X2 | df | p | |||
Male | 547 | 48.07 | 152 | 44.18 | 1.60 | 1 | 0.206 | ||
Grade | 13.3 | 6 | 0.039 | ||||||
Grade 7 | 191 | 16.78 | 49 | 14.24 | |||||
Grade 8 | 154 | 13.53 | 46 | 13.37 | |||||
Grade 9 | 123 | 10.81 | 36 | 10.47 | |||||
Grade 9+ | 1 | 0.09 | 0 | 0 | |||||
Grade 10 | 233 | 20.47 | 94 | 27.33 | |||||
Grade 11 | 233 | 20.47 | 49 | 14.24 | |||||
Grade 12 | 203 | 17.84 | 70 | 20.35 | |||||
Perceived Health Status | 30.52 | 2 | < .001 | ||||||
Good | 509 | 44.73 | 110 | 31.98 | |||||
Average | 577 | 50.70 | 196 | 56.98 | |||||
Poor | 52 | 4.57 | 38 | 11.05 | |||||
Perceived Weight | 23.87 | 2 | < .001 | ||||||
Underweight | 142 | 12.49 | 40 | 11.63 | |||||
Average | 655 | 57.56 | 153 | 44.48 | |||||
Overweight | 341 | 29.96 | 151 | 43.9 | |||||
Academic performance (scores) | 12.60 | 3 | 0.006 | ||||||
85 ~ 100 | 188 | 16.52 | 35 | 10.17 | |||||
75 ~ 85 | 341 | 29.96 | 92 | 26.74 | |||||
60 ~ 75 | 492 | 43.23 | 172 | 50 | |||||
lower than 60 | 117 | 10.28 | 45 | 13.08 | |||||
Perceived academic pressure | 35.24 | 2 | < .001 | ||||||
Light | 58 | 5.10 | 13 | 3.78 | |||||
Average | 726 | 63.8 | 164 | 47.67 | |||||
Heavy | 354 | 31.11 | 167 | 48.55 | |||||
Family Social Economic Status | 8.02 | 2 | 0.018 | ||||||
Rich | 181 | 15.91 | 56 | 16.28 | |||||
Enough | 915 | 80.4 | 263 | 76.45 | |||||
Poor | 42 | 3.69 | 25 | 7.27 | |||||
Parental Marriage Status | 5.13 | 5 | 0.401 | ||||||
Normal | 1,034 | 90.86 | 305 | 88.66 | |||||
Re-marriage | 25 | 2.2 | 9 | 2.62 | |||||
Divorced | 50 | 4.39 | 14 | 4.07 | |||||
Live apart | 18 | 1.58 | 9 | 2.62 | |||||
Widowed | 5 | 0.44 | 2 | 0.58 | |||||
Others | 6 | 0.53 | 5 | 1.45 | |||||
Family Relationship | 17.79 | 2 | < .001 | ||||||
Good | 846 | 74.34 | 220 | 63.95 | |||||
Average | 279 | 24.52 | 113 | 32.85 | |||||
Poor | 13 | 1.14 | 11 | 3.2 | |||||
Relationship with Teacher | 20.61 | 2 | < .001 | ||||||
Good | 467 | 41.04 | 97 | 28.2 | |||||
Average | 649 | 57.03 | 234 | 68.02 | |||||
Poor | 22 | 1.93 | 13 | 3.78 | |||||
Relationship with Classmates | 29.65 | 2 | < .001 | ||||||
Good | 640 | 56.24 | 140 | 40.7 | |||||
Average | 480 | 42.18 | 190 | 55.23 | |||||
Poor | 18 | 1.58 | 14 | 4.07 | |||||
Clinical Feature | Mean | SD | Mean | SD | Z | df | p | ||
CES-D Score | 13.5 | 8.91 | 20.74 | 11.41 | -11.10 | < .001 | |||
WHOQOL Score | |||||||||
Physical Domain | 14.91 | 2.34 | 13.29 | 2.66 | -10.11 | < .001 | |||
Psychological Domain | 14.22 | 2.99 | 12.62 | 3.25 | -7.98 | < .001 | |||
Social Relationship | 15.22 | 3.41 | 13.85 | 3.93 | -5.76 | < .001 | |||
Environmental Domain | 14.45 | 2.79 | 13.62 | 30.3 | -4.62 | < .001 | |||
Overall | 14.51 | 3.24 | 13.01 | 3.54 | -7.17 | < .001 | |||
Internet Addiction Test Score | 35.02 | 7.42 | 61.18 | 10.6 | -28.15 | < .001 | |||
N | % | N | % | X2 | df | p | |||
Met the cut-off for Depressive Emotion | 89 | 7.82 | 83 | 24.13 | 68.47 | 1 | < .001 |
Before analyzing the independent relationship between IA and QoL scores, associations between key continuous variables were illustrated in Table 3. There was an evident negative association between IAT score and student’s self-rated QoL scores in all domains. The results in the simple linear regression analyses revealed that whether the student has met the cutline for IA explained the variances in QoL scores of physical health, psychological health, social relationship and environmental domain by 9.5%, 5.6%, 3.5% and 2.3% respectively.
After adjusted for depressive mood, the multivariate regression analyses demonstrated that IA was still significantly negatively associated with QoL score in every domain. However, when all variables were controlled, only physical health (B= -0.87, p < .001), psychological health (B= -0.40, p = .011) and overall QoL (B= -0.45, p = .006) were negatively associated with IA. The multivariate regression models indicating the association between IA and QoL scores accounted for 31.8% of variance in physical health and 40.1% in psychological health. The regression models were demonstrated in Table 4.
Age | CES-D | IAT | Physical Domain | Psychological Domain | Social relations | Environmental Domain | Overall | |
---|---|---|---|---|---|---|---|---|
Age | 1 | |||||||
CES-D | 0.10** | 1 | ||||||
IAT | 0.09** | 0.39** | 1 | |||||
Physical Domain | -0.13** | -0.57** | -0.38** | 1 | ||||
Psychological Domain | -0.12** | -0.69** | -0.32** | 0.69** | 1 | |||
Social Relations | -0.12** | -0.46** | -0.24** | 0.52** | 0.60** | 1 | ||
Environmental Domain | -0.12** | -0.52** | 0.25** | 0.61** | 0.69** | 0.60** | 1 | |
Overall | -0.20** | -0.52** | 0.29* | 0.61** | 0.61** | 0.48** | 0.53** | 1 |
Note: * indicated p < 0.05; ** indicated p < 0.01. |
Changes in means of QoL scores | ||||
---|---|---|---|---|
Models | β | 95%IC | p | |
Univariate Regression Models | ||||
1a | univariate regression model for association between internet addiction and physical QoL score (all available data, N = 1725) | -1.87 | (-2.14, -1.60) | < .001 |
1b | univariate regression model for association between internet addiction and physical QoL score (complete case, N = 1482) | -1.62 | (-1.91, -1.32) | < .001 |
2a | univariate regression model for association between internet addiction and psychological QoL score (all available data, N = 1725) | -1.75 | (-2.09, -1.41) | < .001 |
2b | univariate regression model for association between internet addiction and psychological QoL score (complete case, N = 1482) | -1.59 | (-1.96, -1.23) | < .001 |
3a | univariate regression model for association between internet addiction and social relation QoL score (all available data, N = 1725) | -1.58 | (-1.97, -1.19) | < .001 |
3b | univariate regression model for association between internet addiction and social relation QoL score (complete case, N = 1482) | -1.36 | (-1.79, -0.94) | < .001 |
4a | univariate regression model for association between internet addiction and environment QoL score (all available data, N = 1725) | -1.05 | (-1.37, -0.73) | < .001 |
4b | univariate regression model for association between internet addiction and environment QoL score (complete case, N = 1482) | -0.83 | (-1.17, -0.48) | < .001 |
5a | univariate regression model for association between internet addiction and overall QoL score (all available data, N = 1725) | -1.78 | (-2.14, -1.41) | < .001 |
5b | univariate regression model for association between internet addiction and overall QoL score (complete case, N = 1482) | -1.50 | (-1.90, -1.10) | < .001 |
Multivariate Regression Models | ||||
1c | Association between Internet Addiction and physical QoL: Adjusted for depressive mooda | -1.25 | (-1.53, -0.96) | < .001 |
2c | Association between Internet Addiction and psychological QoL: Adjusted for depressive mooda | -0.90 | (-1.24, -0.56) | < .001 |
3c | Association between Internet Addiction and social relation QoL: Adjusted for depressive mooda | -0.86 | (-1.28, -0.44) | < .001 |
4c | Association between Internet Addiction and environment QoL: Adjusted for depressive mooda | -0.40 | (-0.73, -0.06) | 0.021 |
5c | Association between Internet Addiction and overall QoL: Adjusted for depressive mooda | -1.02 | (-1.42, -0.63) | < .001 |
1d | Association between Internet Addiction and physical QoL: Adjusted for all confoundersb | -0.87 | (-1.13, -0.60) | < .001 |
2d | Association between Internet Addiction and psychological QoL: Adjusted for all confoundersb | -0.40 | (-0.70, -0.09) | 0.011 |
3d | Association between Internet Addiction and social relation QoL: Adjusted for all confoundersb | -0.36 | (-0.75, 0.02) | 0.063 |
4d | Association between Internet Addiction and environmentalQoL: Adjusted for all confoundersb | 0.02 | (-0.29, 0.33) | 0.906 |
5d | Association between Internet Addiction and overall QoL: Adjusted for all confoundersb | -0.45 | (-0.78, -0.13) | 0.006 |
Note: a. The ‘depression mood’ adjusted in the regression models refers to ‘with or without depression mood (i.e. CES-D ≥ 28)’ rather than CES-D scores; b. Confounders adjusted in the models include: age, gender, grade, self-perceived health status, self-perceived weight, self-perceived academic performance, self-perceived pressure of study, relationship with classmates, relationship with teacher, parental marriage status, family relationship, SES and with/without depressive mood (i.e. CES-D ≥ 28). |
Many studies have illustrated the prevalence of IA in Chinese adolescents and risk factors prone to IA as well as its impact on psychological wellbeing. However, fewer studies have underpinned the pure relations between QoL and IA. The current study attempted to reveal the prevalence of IA in Shenzhen and the unique associations between IA and four different domains of QoL, eliminating effects of confounding factors. The prevalence of IA was 23.2% in the current sample, which is around the same rate as in Hong Kong (17–26.8%) [38], but higher than global prevalence (6%) [14]. The concurrence of IA and depressive mood was common. 24.13% of the students with IA reported depressive mood, whereas only 7.82% of those without IA did. Moreover, students with IA tended to have lower QoL scores in all of the four domains. Even after controlling important risk factors, especially the depressive mood, IA was still significantly in relation to lower subjective QoL in terms of physical health, psychological health and overall perception.
Given that different assessment tools and cut-off values were used in different studies, discrepancies in prevalence of IA lay in previous literature. When compared with studies using same scale and cutoff line, the prevalence of IA in Shenzhen was higher than other parts of Mainland China including Guanzhou (10.8%) [39], Anhui (12.0%) and Qinghai (15.0%) [25]. This was inconsistent with previous studies suggesting that economic developed areas had lower IA prevalence than developing areas. On the other hand, stress-eliciting environment is suggested to increase the rate of IA [40], since people tend to engage in Internet for psychological relief. Although as an economic developed area in China, Shenzhen have more access to other outdoor activities for stress relief, greater academic pressure was also seen in Shenzhen middle schools and high schools, which contributed to the high rate in IA [41]. In the current sample, 34.5% of the students reported great pressure on academic study, higher than those in Qinghai and Anhui [25]. With high academic pressure, students tend to spend more time indoor, limiting their access to recreational activities and Internet was convenient to approach. Also, the access to Internet increased rapidly in the recent two decades. Devices like mobile phones and tablets, as well as increased accessibility to Wi-Fi provided easy access to Internet. Coupled with the fact that Shenzhen is famous for its IT industry, the more convenient access to electronic devices might trigger the increase in IA [4].
In the line with previous findings, IA is significantly associated with depressive mood. Current study suggested that students with IA tended to have a higher mean CES-D score and were more likely to meet the criteria for depressive mood. Although different scales were used to classify depressive mood, a systematic review of 20 studies published before 2014 showed that three quarters of studies supports a significant association between IA and depression [42]. A recent study by Lu and colleagues in 2018 also used CES-D to assess the depressive level in Chinese adolescents [25]. They suggested that students with depressive mood were more likely to engage with IA. Our results reconfirmed that the risk of engaging with IA is more significant if the student is currently in a depressive mood.
The mood enhancement hypothesis put forwarded by Bryant and Zillmann argued that people tend to engage with leisure actives (e.g. internet surfing, watching TV) than other activities when they are with negative emotions [26]. Also, Internet use could somehow compensate for the lack of social engagement [25]. On the other hand, social displacement hypothesis stated that indulging in online social communications results in decrease in real life social interaction, which leads to maladjustment problems [17]. When online socialization displaces offline social communication, it is likely to lead to the emergence of negative emotions, especially in adolescents [43]. Although the current study did not specify the causal relationship between depression and IA, it is consistent with previous literatures and indicated a positive relationship between depressive mood and IA. Depressive status leads to social withdrawal and decreases people’s interest to activities that they used to enjoy, eventually increasing risk of overindulging in Internet use. In turn, IA increases the risk of developing a depressive mood, leading to a viscous cycle of constantly having depressive mood and IA.
The results in univariate regression models clearly indicated negative associations between IA and QoL in all aspects. Such associations remained significant even after controlling for the existence of depressive mood, a factor that holds a close relationship with lower physical and psychological satisfaction and declines in social relation [44]. Our results suggested that IA, independent from depressive mood, may also increase the risk of lower QoL in terms of physical and psychological wellbeing, as well as satisfaction in social relation and environment in some degree. Such negative associations identified between IA and QoL, thereby, support the I-PACE that excessive Internet use eventually affect daily life physically and psychologically [27].
Excessive time spent on Internet reduced the time for outdoor physical activities and face to face socialization, resulting in a lower QoL rating for physical health. Reduction in outdoor activities and physical unwellness brought by long screen time led to lower satisfaction on physical health [45]. In terms of psychological health, depressive mood is believed to significantly account for lower self-rated score in psychological health. However, our result suggested that IA itself also account for low QoL score for psychological health, though the multitude of the effect was small. However, the negative association between IA and QoL in social relation domain became insignificant after controlling all other confounders (e.g. relations towards important people in life). As the WHOQoL rating scale is a generic rating scale, it might not be sensitive enough to discern the changes in social relationship. Therefore, after controlling the effect of important relationships in students’ life, the association between IA and QoL in social relation domain reduced. Yet, this is consistent with the idea that real life relationship plays an important role in life satisfaction towards social relation domain [46]. The social displacement hypothesis proposed that excessive online communication may interrupt the establishment in real life relationship [17]. Though our result did not suggest a pure relationship between QoL in social relation and IA, there is a potential that IA affects the satisfaction in social relation for future discussion. Thus, the result again provided evidence to the negative impact of IA on psychological satisfaction to some extent.
Since Internet is essential in modern society, it is unrealistic to completely prohibit Internet from young people. Therefore, in order to improve the QoL of adolescents, it is important to manage IA with a risk-focused approach [47]. A risk-focused approach refers to developing interventions for conditions with a focus on their related factors. Following the results that IA is associated with QoL in physical and psychological domains, intervention for IA should stress on engaging them in physical exercises, enhancing coping strategies for psychological stress and developing new skills to improve psychological health. School and caregivers should be engaged in early screening of depressive mood signs and provide in-time interventions, reducing the risk of engaging with IA. Also, with the consideration in gender differences, skills for constructive use of Internet and balancing online and offline social time are useful for female and mood regulation skills training are for males to minimize the impact of IA [43].
One of the limitations was the missing data induced when adjusting confounding variables, resulting in concerns in interpreting the results. Firstly, the sample size was shrunk from 2148 to 1418 in completed case sample, which may reduce the statistical power of the associations between IA and different domains of QoL. Evident association between IA and different domains of QoL seen in univariate models with all available data were weakened in complete case sample, suggesting that missing data reduces the statistical power of the targeted associations. Moreover, current study is a cross-sectional study, in which data were collected at the same period of time. Therefore, the causal relationship between IA and QoL could not be illustrated. Longitudinal studies are called for investigating the causal relationship between IA and QoL to enhance the development of interventions for IA. Thirdly, a mediation analysis is needed to further underlie the role of depressive mood in the association between IA and QoL. Depressive mood itself might be developed as the consequence of IA [48] and it also highly impacts the life quality of students, especially their psychology health. Future study can further investigate the role of depression in a longitudinal study and to better understand the impact of IA, helping with developing interventions.
Despite of the limitations, the current cross-sectional study investigated the prevalence of Internet Addiction and its association with QoL in Chinese Adolescents independent from depression and other demographic factors. It suggested that adolescents with IA are more likely to have a lower perception of their QoL in physical health and psychological health. IA itself also affects QoL in social relationship domains in some degree. Therefore, interventions to manage and prevent Internet Addiction are needed, especially with a focus on physical health and psychological health.
Internet Addition
the Fifth Edition of Diagnostic and Statistical Manual of Mental Disorders
Quality of Life
Attention Deficit Hyperactivity Disorder
Person-Affect-Cognition-Executive model
Disability-adjusted life year
Healthy life expectancy
Great Bay Area
Internet Addiction Test
Centre for Epidemiologic studies of Depression Symptom Scale
Community Mental Health Assessment
World Health Organization Quality of Life
Analysis of Covariance
Ethics Approval and consent to participate: The consent form was approved by Shenzhen Kangning Hospital Medical Ethics Approval Committee in December, 2016. Participants received questionnaires with an attached consent form and brought back to their parents/guardians. Both participants and their parents/guardians were informed about the study and signed the consent form to show their agreement for using their data in the current study.
Consent for Publication: Not Applicable.
Availability of data and material: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing Interests: Not Applicable.
Funding: Sanming project of medicine in Shenzhen (No. SZSM201612079) and Shenzhen Double Chain Grant [2018]256
Authors’ contribution: Miss Siying Wen and Dr. Xiaolan Cao contributed equally in generating and designing the study. This manuscript was written up by Miss Siying Wen mainly, and was amended by Dr. Xiaolan Cao and Dr. Jianchang Xu. Dr. Jianping Lu supervised the whole project in terms of its design and conduction, provided suggestions for the write-up and finalized the manuscript.
Acknowledgement: This study was funded by Sanming project of medicine in Shenzhen (No. SZSM201612079) and Shenzhen Double Chain Grant [2018]256. The authors would like to thank Shenzhen Luohu District Bureau of Education and Shenzhen Futian District Bureau of Education for their tremendous support.